Abstract

Microplastics pose a serious risk to biodiversity in the oceans and to global public health. Detecting microplastics in ocean water can be time-consuming and labor-intensive, limiting the sorts of ongoing assessments one can complete, and it is especially true of recent techniques like spectroscopy and microscopy. A method outline is proposed in which microplastics can be detected in real time, and the automatic classification of microplastics is achieved using machine learning. The design involves AI-enhanced optical sensors and IoT devices that collect the data. It recommends real-time in situ particle detection using fluorescence optical sensors and classification of spectro-morphological dendritic microplastics using a Convolutional Neural Network (CNN). It detects and classifies microplastics while also telling the difference between organic and inorganic particulates. Real-time data and visual analytics, using a cloud-based IoT platform, can be used for pollution forecasting and environmental Monitoring. The methodology proposes validated field work in both freshwater and estuarine environments, which resulted in an average classification accuracy of 95.8 percent and real-time processing latency of less than 2.3 seconds per sample. Using advanced sensors to combine AI and IoT provides the ability to monitor microplastics in real time as a scalable solution. It enables the active control of contaminated water and the preservation of water bodies.

Keywords: Microplastics detection, Monitoring in real time, Machine learning, Optical sensing, Water bodies, Environmental monitoring